Computational intelligence modeling of nanomedicine preparation using advanced processing: Solubility of fludrocortisone acetate in supercritical carbon dioxide

Abstract

The method of green technology which is based on supercritical solvent has been studied in this work for analysis of nanomedicine preparation of solid dosage oral medications. Given that the poor drug solubility in aqueous media is a big challenge in pharmaceutical industry, nanomedicines would help improve the drug solubility in aqueous media. The solubility of fludrocortisone acetate in supercritical carbon dioxide is modelled in this research using various machine learning methods because it is a crucial aspect of the expansion of the pharmaceutical business. For this purpose, the accessible data have two input features: a pressure range of 120–300 (bar) and a temperature range of 308–338 (K). MLP, v-SVR and MLR are the basic models used in this research, but not their raw versions. They are improved for modeling drug solubility and coupled with the grey wolf optimization (GWO) in order to optimize the models. The models optimized by GWO showed acceptable results, but among these models, MLP regression has shown better results when coupled with this optimization algorithm. This model has the RMSE error rate of 2.98 × 10−2 and its R2 score is 0.9797 in correlating the solubility data of the medicine.

Keywords

• Nanomedicine
• Multilayer perceptron
• Support vector machine
• Multi linear regression
• Drug solubility